首页|临床预测模型诊断新生儿坏死性小肠结肠炎的研究进展

临床预测模型诊断新生儿坏死性小肠结肠炎的研究进展

扫码查看
新生儿坏死性小肠结肠炎(NEC)是一种常发生在早产儿和低出生体重儿的严重肠道疾病,延迟治疗可能会发生神经发育迟缓、生长发育不良、胃肠道狭窄、短肠综合征、肠道功能衰竭等严重并发症,早期诊断及干预有助于改善其预后.早产、肠内喂养、肠道菌群异常定植、肠道黏膜缺血、感染等都是NEC发生的主要危险因素,但具体发病机制尚未完全明确.临床预测模型(CPMs)是利用数学公式估计当前个体可能患有某病或将来发生某结局的概率模型.近年来,通过临床特征、实验室指标、生物学标志物、微生物等信息构建的CPMs对NEC的疾病诊断发挥着重要作用.然而,研究结果却不相一致,缺乏对NEC诊断预测模型的总结概括以推动临床实践.本文就传统统计方法和机器学习(ML)所构建的NEC诊断模型展开综述.
Research Progress of Clinical Prediction Models in Diagnosis of Neonatal Necrotizing Enterocolitis
Neonatal necrotizing enterocolitis(NEC)is a serious intestinal disease that often occurs in preterm and low birth weight infants,and delayed treatment may lead to serious complications such as neurodevelopmental delay,poor growth and development,gastrointestinal stricture,short bowel syndrome,intestinal failure,early diagnosis and intervention can help improve its prognosis.Preterm birth,enteral feeding,abnormal colonization of intestinal flora,intestinal mucosal ischemia,and infection are the main risk factors for the occurrence of NEC,but the specific pathogenesis is not fully understood.Clinical prediction models(CPMs)is a probability model that uses mathematical formulas to estimate the current individual may have a disease or a future outcome.In recent years,CPMs constructed by clinical features,laboratory indicators,biological markers,microorganisms and other information plays an important role in the diagnosis of NEC.However,the results of the study are inconsistent,and there is a lack of summary generalizations of the NEC diagnostic predictive model to advance clinical practice.This paper reviews the NEC diagnostic models constructed by traditional statistical methods and machine learning(ML).

Neonatal necrotizing enterocolitisClinical prediction modelsTraditional statistical methodsMachine learning

刘代成、王化彬、任雪云

展开 >

济宁医学院 山东 济宁 272000

济宁医学院附属医院儿科 山东 济宁 272000

新生儿坏死性小肠结肠炎 临床预测模型 传统统计方法 机器学习

2023年山东省自然基金青年项目山东省博士后创新人才支持计划项目济宁医学院2021年高层次科研项目培育计划

ZR2023QH382SDBX2022020JYGC2021FKJ010

2024

中国医学创新
中国保健协会

中国医学创新

影响因子:1.706
ISSN:1674-4985
年,卷(期):2024.21(23)